Semantic network model in Chinese vocabulary generation and evolution

Abstract

Abstract This study delves into the semantics of Chinese vocabulary generation and evolution, providing insights into the profound connotations of the Chinese language and culture. We begin by vectorizing Chinese vocabulary data using the Word2Vec algorithm, which is subsequently input into a deep neural network model. Within this framework, the neurons are tasked with extracting semantic features pertinent to the developmental trajectory of Chinese vocabulary. Following this, the Intersection over Union (IoU) value for each semantic feature is computed, enabling the semantic annotation of the vocabulary. Additionally, the formulation of constraint rules significantly enhances the model’s accuracy in semantic recognition analysis. This approach not only advances our understanding of linguistic structures but also enriches our appreciation of the cultural dynamics embedded within the language. The accuracy and recall of the semantic network model in this study are much higher than those of other comparative models, with an average semantic annotation accuracy of 91.33% in the three Chinese vocabulary datasets, which supports further research and analysis in follow-up. The exploration of the practical application of the semantic network model shows that it achieves better results in the semantic identification and annotation of words in poems, and it is found that the lexical semantic class with the highest frequency in the Three Hundred Song Lyrics is {Func word}, which accounts for 0.056. The semantic network model designed in this paper provides an effective method in the field of Chinese vocabulary semantics analysis and lays the foundation for a deep understanding of the generation and evolution of the culture of the Chinese language. A foundation is laid for a deep understanding of the generation and evolution of Chinese language culture.